University of Southern California Dissertations and Theses

Los Angeles County vote center site-selection: facilitating decision making with a web application

(USC Thesis Other)

Los Angeles County vote center site-selection: facilitating decision making with a web application

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Los Angeles County Vote Center Site-Selection: Facilitating Decision Making with a Web Application

by

Alexander Holt

A Thesis Presented to the
Faculty of the USC Graduate School
University of Southern California
In Partial Fulfillment of the
Requirements for the Degree
Master of Science
(Geographic Information Science and Technology)

May 2019

Copyright © 2018 by Alexander Holt

To my Mother, Father, and Ashley-Rose

iv
Table of Contents
List of Figures ................................................................................................................................ vi
Acknowledgements ....................................................................................................................... vii
List of Abbreviations ................................................................................................................... viii
Abstract .......................................................................................................................................... ix
Chapter 1 Introduction .................................................................................................................... 1
1.1. Progression of Election Systems .........................................................................................1
1.2. Senate Bill 450: Vote by Mail Voting and Ballot Discussions ...........................................2
1.3. Motivations and Project Goals ............................................................................................8
Chapter 2 Related Work................................................................................................................ 10
2.1. Voting Locations effect on Voting Decisions ...................................................................10
2.2. Constituency and the Voting Process in Los Angeles County .........................................11
2.3. Voter Participation and Turnout .......................................................................................12
2.4. Other Jurisdictional Approaches to Applying the Voter’s Choice Act.............................12
2.5. Other LA County Projects Related to VSAP ....................................................................17
Chapter 3 Data and Methodology ................................................................................................. 19
3.1. Web Application Design and Structure ............................................................................19
3.2. Potential Vote Center Site List .........................................................................................23
3.3. Voter Registration and Turnout Toolset ...........................................................................26
3.4. Potential Vote Center Service Layer through Vehicle Accessibility ................................30
3.5. Initial User Feedback ........................................................................................................34
Chapter 4 Results .......................................................................................................................... 35
4.1. Overview of Functionality for the Application .................................................................35
4.2. Application in Use ............................................................................................................43
4.3. Feedback from reviewers ..................................................................................................47
v
Chapter 5 Future Work and Conclusions ...................................................................................... 50
5.1. Current Status of the Application ....................................................................................50
5.2. Future Improvements to the Application and Fixing Bottlenecks ....................................51
5.3. Final Findings ...................................................................................................................54
5.4. Conclusion ........................................................................................................................55
References ..................................................................................................................................... 56

vi
List of Figures
Figure 1 LA County Infographic Summarizing Changes for SB 450 ............................................ 3
Figure 2 LA County’s VSAP Project Time line ............................................................................. 5
Figure 3 Los Angeles County New Voting System Prototype ....................................................... 7
Figure 4 Percentage of Eligible Residents Not Registered to Vote in Sacramento County ......... 15
Figure 5 Priority Ranking Map Sacramento County .................................................................... 16
Figure 6 CCEP Vote Center Siting Tool....................................................................................... 17
Figure 7 SQL Query through Registered Voter Database to Acquire Voters by Age Class ........ 29
Figure 8 Splash Screen Display .................................................................................................... 36
Figure 9 Default Layers of the Web GIS Application .................................................................. 37
Figure 10 Icons of the Main Tools ................................................................................................ 37
Figure 11 Demonstration of the “Filter by City/Community” Tool ............................................. 38
Figure 12 “Select Vote Center Location Type” Tool ................................................................... 39
Figure 13 Selectable Registered Voter and Election Turnout Layers ........................................... 40
Figure 14 Registered Voters by Age Class 18-29 Relative to the Countywide Mean .................. 40
Figure 15 General Election 2016 Turnout by Registered Voter ................................................... 41
Figure 16 Opening the Service Area Site Tool and Entering the Service Area ID ...................... 42
Figure 17 Running the Service Area Site Tool Displaying a 5-minute Drive Polygon, Voter
count, and Business Count ............................................................................................................ 42
Figure 18 Demonstration of Selecting an Election Precinct with a Younger Age Demographic of
Registered Voters. ......................................................................................................................... 45
Figure 19 Inglewood Continuation High School Service Area .................................................... 46
Figure 20 Inglewood High School Service Area .......................................................................... 46
vii

Acknowledgements
I would like to thank my thesis advisor Assistant Professor and Director of Graduate Studies, Dr.
Robert Vos of the Dana and David Dornsife College at the University of Southern California. Dr.
Vos never gave up on me when I struggled to define my topic or worried about being able to
make enough time in my schedule to complete the thesis. He pushed me through the difficult
times by meeting with me every week, giving me countless hours of thoughtful insight, and
helping direct my thesis into a product that I feel very proud of. I would also like to thank the
GIS staff of the Los Angeles County Registrar-Recorder Office for giving me direction on how
to aggregate the large voter datasets that were used in this thesis and letting me have access to all
the tools that were required to complete the web GIS application.

viii
List of Abbreviations
AdCom Advisory Committee (for the VSAP project in LA County)
CCEP California Civil Engagement Project
CSV Comma-separated values
Esri Environmental Systems Research Institute
GIS Geographic information system
LA County Los Angeles County
RRCC Los Angeles County Registrar-Recorder Office
SB 450 Senate Bill No. 450: California Voter’s Choice Act
SSI Spatial Sciences Institute
UCLA University of California Los Angeles
USC University of Southern California
VBM Vote by Mail
VCA California Voter’s Choice Act
VSAP Voting Solutions for All People Project

ix
Abstract
Elections are a pinnacle of modern democracy. Fair elections have been a key pillar to the
success of representative democracy; yet, many of the age-old problems still exist. One of these
problems is voter turnout and making voting accessible to every member of that democracy. In
today’s elections, we are at an interesting crossroads of technology and convenience. California
has recognized both trends and is implementing the Voter’s Choice Act (VCA) or SB 450 to
guide the future of voting in the state by allowing people to vote over a multi-day period and at
any location of their choosing. With this change, new voting locations will play a more critical
role to the process than ever before. This paper discusses the design and implementation of a
web GIS application that will assist Los Angeles County stakeholders and elections officials in
choosing these new voting locations by providing tools for analysis. The tools developed are
designed to assess potential site locations and their relationship to voters and businesses, voter
age demographics, and voter turnout. The initial version of web GIS application documented
here was reviewed by county officials. It will likely be provided to local stakeholders by the Los
Angeles County Registrar-Recorder office and used in the first round of site-selection for these
new voting locations.

1

Chapter 1 Introduction
The purpose of this thesis is to assist in determining vote center locations for Los Angeles
County’s new voting system through a multi-step spatial analysis process. Currently the Los
Angeles County Registrar-Recorder/County Clerk office is undergoing a redesign of their entire
voting system called the “Voting Systems for All People” or VSAP. Under a new 2016
California Senate Bill titled “SB-450 Elections: Vote by Mail Voting and Mail Ballot Elections”
or the “California Voter’s Choice Act,” many new requirements were created to equitably
distribute voting locations across all voting areas. This thesis conducts various spatial analysis
techniques and presents the results in a Web GIS application to assist in accomplishing the goals
of this bill in choosing new vote center locations for Los Angeles County.
1.1. Progression of Election Systems
Elections within the United States have had a dramatic evolution over the last two
hundred years, since the inception of one of the world’s oldest and most stable democracies. In
coordination with rapidly evolving technology of the 21
st
century, new election systems are
being tested and implemented. Figuring out the best way to apply these new technologies in
cohesion with new rules and regulations, as well as appeasing all participants in the system is a
long process that requires constant input. Understanding past technologies and the progression of
the elections is essential in being able to determine the best path forward.
The progression of voting technology has had a slow evolution due to sensitivity of
elections and their role in politics. Before the 1800s voting was typically done by voice, in a
public setting (Troy 2016). Paper ballots, a variation of the technology most of the country uses
today, eventually got phased in as the American government grew and political factions rose to
power. Paper ballots were described by King (2016) as being, “…easy to count and hard to

2

falsify, but the system was better suited to smaller elections” (34). Not until the 19
th
century,
when controversy in American politics began to rise due to the civil war and subsequent
limitations on voting rights, did the voting process enter a stage of secrecy. The term “vest-
pocket” was coined for people who would show up to vote with the ballots hidden, which in turn
also created mass voting fraud by false ballots. As the number of technical issues grew, new
technological solutions were developed to thwart voting fraud. Lever machines were created to
help count ballots in 1920, and in 1965 the punch card technology was invented (McGerr 1984).
This was the general technology used for the latter half of the 20
th
century. A variation of that
voting machine called the “Votomatic” is still in use in a few counties today. The progress at
which both regulation and legislation govern voting equipment has been slow to keep up with
modern technology. This in turn has limited access to voting and voter turnout has suffered.
1.2. Senate Bill 450: Vote by Mail Voting and Ballot Discussions
To improve election systems, in 2016 California passed Senate Bill No. 450 titled “Vote
by Mail Voting and Mail Ballot Elections,” written by state Senators Ben Allen and Robert
Hertzberg. The bill was specifically designed to guide and improve the implementation of
elections conducted in California. The bill was made up of four parts: (1)requiring all counties to
mail registered voters a VBM (vote-by-mail) ballot; (2)replacing the regular polling places with
a new type of voting location called a vote center; (3)creating a new time window of ten days
before the election during which voters can vote; (4)and requiring every jurisdiction to create its
own plan to implement these new changes while receiving input from the public (Los Angeles
County 2017). The final bill was passed into a law on September 29, 2016, outlining the
framework for jurisdictions to implement an entirely new voting system.

3

Figure 1 LA County Infographic Summarizing Changes for SB 450 (LA County 2017)
Figure 1 summarizes many of the critical changes SB 450 will have on Los Angeles
County’s current voting system. One of the major components of the bill is to open the time
frame for voting. Instead of elections being on a single day and each voter having a specific poll
location to go to, voters will have ten days and be able to use any vote center. The official
terminology of the location where participants are going to vote is also being changed from
“polling place” to “vote center.”
The primary objective of the bill is for technological innovation to improve the voting
process and create optionality for larger counties like Los Angeles County, by laying out an

4

easier framework on how elections should be conducted. As the law itself states, “This bill
would, on or after January 1, 2020, authorize the County of Los Angeles to conduct any election
as a vote center election if certain conditions are satisfied, including conditions related to ballot
drop-off locations and vote centers” (SB 450 2016). This verbiage allows LA County to work
with local elected officials to create a proper set of conditions to be used in deciding the location
of vote centers and drop-off locations. The following is a minimum set of conditions that the bill
requires:
The county elections official, when developing the draft plan for the administration of
elections conducted pursuant to this section, considers, at a minimum, all of the
following:
(i) Vote center and ballot drop-off location proximity to public transportation.
(ii) Vote center and ballot drop-off location proximity to communities with
historically low vote by mail usage.
(iii) Vote center and ballot drop-off location proximity to population centers.
(iv) Vote center and ballot drop-off location proximity to language minority
communities.
(v) Vote center and ballot drop-off location proximity to voters with disabilities.
(vi) Vote center and ballot drop-off location proximity to communities with low rates
of household vehicle ownership.
(vii) Vote center and ballot drop-off location proximity to low-income communities.
(viii) Vote center and ballot drop-off location proximity to communities of eligible
voters who are not registered to vote and may need access to same day voter
registration.
(ix) Vote center and ballot drop-off location proximity to geographically isolated
populations, including Native American reservations.
(x) Access to accessible and free parking at vote centers and ballot drop-off
locations.
(xi) The distance and time a voter must travel by car or public transportation to a
vote center and ballot drop-off location.
(xii) The need for alternate methods for voters with disabilities for whom vote by
mail ballots are not accessible to cast a ballot.
(xiii) Traffic patterns near vote centers and ballot drop-off locations.
(xiv) The need for mobile vote centers in addition to the number of vote centers
established pursuant to this section (SB 450 2016).

These conditions require an extensive site selection process that Los Angeles County is currently
undertaking. The spatial web application developed through this project will allow local

5

stakeholders to assist and guide the site selection process. Spatial analysis and the development
of a proper site selection system is essential for helping meet all these conditions and fulfilling
the purpose of SB 450.
The passing of SB 450 prompted Los Angeles County to go on a decade-long journey to
optimize the entire voting system through a project titled “Voting Systems for All People” or
VSAP. The current system, which was originally implemented in the 1960s, is aging, the cost to
replace and maintain the old equipment is increasing, and the efficiency/performance of the
entire operation is deteriorating. These factors prompted the head of Los Angeles County’s
Registrar-Recorder’s Office, Dean Logan, to initiate a five-phase project in September 2009 as
shown in Figure 2 below (Los Angeles County 2017).

Figure 2 LA County’s VSAP Project Time line (LA County 2017)

6

Phase I and Phase II were the preliminary stages of the project. Election stakeholders
from different groups with various expertise were chosen to provide input. The stakeholders are
identified as: “…. voters, poll workers, advocates, key community organizations and elections
staff” (Logan 2017). Research was then conducted to evaluate the current election system and
receive feedback from the various stakeholders. This data was compiled, organized, and used to
guide the second phase of the project. Phase II took a deeper dive into the main concerns that
were addressed by these stakeholders by creating a VSAP Advisory Committee or AdCom,
designed to guide the process of creating a new voting system and ensure the “voice of the voter”
was respected throughout its implementation. The two main hurdles in this part of the project
were assessing the voter systems markets and the regulatory environment (Logan 2017). A
research team from UCLA Luskin School of Public Affairs played a key role in the
implementation of Phase II, by providing research on which systems would meet all the goals
that were set in Phase I and outlining the key regulatory problems that would need to be
overcome to reach those goals. The new system would have to be designed with the legislature to
help accommodate all these needs, thus Los Angeles County played a prominent role in the
creation of SB 450. Phase III set out to fulfill all the request that were outlined in the first two
phases and move from research to development and design.

7

Figure 3 Los Angeles County New Voting System Prototype
The voting system prototype was designed by an award-winning global design team named
IDEO. The idea behind this voting system was to allow easy access to every resident of Los
Angeles County whether English was a second language, a voter had a major disability, or
getting to the polls was an issue. Creating easier accessibility to voting and easing the process are
major drivers for the VSAP project, which focuses on the voting procedure itself. From 2012 to
2016 this voting system was critiqued through the steps outlined in Phase III. The outcome of
this process was the new prototype shown in Figure 3 above. The new mobile and easy to use
system was designed to change and vastly improve the interaction of voting, thwarting problems
of the past. Phase IV is where the process sits as of 2018. Deciding where to locate these new
voting systems within vote centers is the next major task of the VSAP Project.

8

1.3. Motivations and Project Goals
The discussion surrounding the creation of a better voting process has been a hotly-
contested one and become even more relevant as of late with the greater polarization the U.S. has
gone through. To create the best, most fair election process possible, an entire new voting
structure had to be created, and this led to the VSAP project.
After a very chaotic 2016 Presidential election, the public eye has become keen on
watching every election. Hackers, voter suppression, election fraud, and an increasing number of
other issues are creating new problems on top of the existing ones within our election process.
This is causing municipalities of all sizes to move ever quicker to calm these worries (Wines
2017). Generally, the strategy is to show that election officials are addressing all potential
threats, but coming up with solutions to problems that are consistently arising in the face of an
ever-changing world is difficult.
Los Angeles County is one of the largest and most diverse voting populations in the
country and has many contentious issues that derive from its sheer size (Census 2016). Ensuring
that all residents have an equal right to vote in a fair and secure election is at the pinnacle of its
democracy. The many immigrants that live in Los Angeles County have moved here to be a part
of a better system. America prides itself on having a long standing and balanced democracy,
which is becoming increasingly threatened as the integrity of voting is questioned. Improving
this process and making sure it is always at its best is something that must be done to ensure the
continued success of our country.
Spatial analysis on elections is wide spread, conducted on many scales, and essential in
figuring out trends. Even the smallest details can help with understanding the results of an
election. This all starts with the casting of each ballot. Figuring out the details and story of the

9

voting process using the tools GIS offers can give an interesting insight into what has the biggest
impact on these elections. These tools can offer insight for decision makers and those who will
be playing integral roles in the VSAP project. A custom web GIS application with special tools
to address the needs listed out in SB 450 will be a great addition to the immense amount of work
that has already gone into Los Angeles County’s VSAP project. This thesis will discuss the
creation of that web application, the data used to create the custom tools used in the application,
and the functionality of the final product.

10
Chapter 2 Related Work
To improve the current election systems, California passed Senate Bill No. 450 titled, “Vote by
Mail Voting and Mail Ballot Elections,” written by state Senators Ben Allen and Robert
Hertzberg in 2016. The bill was specifically designed to guide and improve the implementation
of elections conducted in California, and the law is being implemented county by county
throughout the state. Comparing Los Angeles County to others and using studies of other large
counties that have implemented similar voting procedures can prove to be valuable in assessing
how to implement the new vote center system. Since the implementation of SB 450, many
counties in the state have begun planning to run elections following the requirements of this
mandate. According to SB 450, 14 counties are currently allowed to conduct elections under SB
450 rules. These include Calaveras, Inyo, Madera, Napa, Nevada, Orange, Sacramento, San Luis
Obispo, San Mateo, Santa Clara, Shasta, Sierra, Sutter, and Tuolumne. The rest of California
must practice elections under the Voter’s Choice Act by 2020 (Padilla 2018). Los Angeles
County is planning to have completed the implementation of the Voter’s Choice Act by 2020.
Understanding the importance of both the process of voting and the Voter’s Choice Act
are essential in making it as fair and organized as possible. Whether voters participate and the act
of voting itself can be influenced by the location where citizens vote. Previous studies discussing
these influences need to be studied by those making the decisions on where polling locations will
be.
2.1. Voting Locations effect on Voting Decisions
The designation of voting locations and the requirement of having a specific polling
location assignment for each voter can influence voter’s decisions. Berger et al. (2008) analyzed

11

this phenomenon. Situational context is how environmental contexts influence the decisions and
choices people make. This is especially relevant in the process of voting, which is one critical
choice people make in a democracy. When people are assigned to a polling location, the type of
polling location can influence how they vote. One example of this would be if a voter is assigned
to a school, the voter could be biased to vote for school-related funding initiatives (Berger et. al
2008).
The study is especially relevant to SB 450 because the bill abolishes designated voting
locations by creating the vote center (in which all voters can vote at any location) and opening
vote by mail to the entire constituency. The vote center location designations should be delegated
in a way to prevent these situational bias scenarios, even subtle ones. This is inevitably
impossible in all scenarios, but probable when there are equally suitable voting locations offered
as options to voters. For example, offering both a church and a public library as equally suitable
voting center options for citizens in each city might at least randomize the influence that would
occur on ballot items related to library funding or religious freedom. This is where GIS becomes
valuable by helping to identify a wide variety of possible sites to select for polling locations.
Various GIS systems can be deployed to give decision makers a valuable tool to guide the vote
center location decisions.

2.2. Constituency and the Voting Process in Los Angeles County
Los Angeles County currently uses the InkaVote voting system, in which voters
physically mark ballots for each contest during the election. Not only is this process tedious, it
creates problems in counting votes. In some election races, ballots are not counted, these are
called a “residual vote.” There are many reasons ballots turn into residual votes, including a

12

voter selecting too many candidates or not filling in the ballot in an appropriate amount of time.
In a 2004 Caltech study based on Los Angeles County, the research concluded that voters with
lower educational backgrounds and who speak English as a second language tend to cast more
residual votes (Sinclair 2004). In a county as diverse as Los Angeles County, having fewer
legitimate ballots can have huge impacts on elections, especially in smaller elections. This also
makes certain populations within a constituency feel under represented knowing that these
problems exist. Eliminating these problems and easing the access to vote is the best way to help
engage communities that have felt underserved by the problems of residual votes.
The Los Angeles County VSAP project will reduce these residual ballot issues. The
creation of a simpler and more effective voting system, through electronic booths prevents
residual ballots and eliminates the possibility of having ballots cast improperly. This in turn can
create a better voter turnout and more accurate elections, especially for voters from different
linguistic backgrounds. This is an important aspect that can be addressed with GIS, by laying out
where minorities groups are, and areas that have had significant residual voting problems in the
past. Using GIS, regions with residual votes can be identified and decision makers will evaluate
these regions when deciding where to put the vote center locations. Through time, these
problems can be reduced with new technologies implemented by the VSAP project and properly
locating vote centers.
2.3. Voter Participation and Turnout
The backbone of SB 450 and its main purpose is to increase voter turnout. Voter turnout
is a problem that is becoming increasingly prevalent in elections. According to the Pew Research
Center in a 2018 report, the U.S falls behind most developed countries in voter turnout. As a
representative democracy since its founding, the U.S. needs to have an active voting

13

constituency, especially Los Angeles County with its diversity and high population density.
Bringing in new technology and changing the process of where voters will vote are the two best
ways to achieve a better voter turnout and are the two most important themes of Los Angeles
County’s VSAP project.
Participation and accessibility typically relate to the distance a voter must travel to his
designated voting location. Removing voting designations eases the travel distance by creating a
more flexible system by allowing voters to choose where to go vote. A 2003 study by J.G
Gimpel titled, “Political participation and the accessibility of the ballot box” concluded that the
physical distance a voter travels has a direct effect on the voter’s likelihood of voting. The study
also states that making the voting location as accommodative as possible will also increase voter
participation. This includes following compliance with the Americans with Disabilities Act
(ADA), having enough space at the voting locations for voting machines, and assessing traffic
and road conditions. The conclusion and overall theme of the article states that physical
accessibility to a polling place is one of the major factors that contributes to voter turnout.
One of the more prolific books written about elections titled “Contemporary
Democracies” by G. Bingham Powell (1982), is a frequently referenced literature among
academics that study election systems and voter turnout. The importance of his work centers on
trends that affect voter turnout. At the crux of the discussion on voter turnout, Powell discusses
the influence of the social and economic environment. A paper on this topic by Blais and
Dobrzynska (2008) discusses how the socioeconomic environment influences turnout, by stating
that there is significant evidence that shows turnout in economically prosperous areas will be
higher. This influence has been downplayed through the years, but significant evidence in recent
studies show how prevalent it really is. Blais and Dobrzynska (2008) state that it relates to all

14

scales of societies from countries all the way down to the municipal level. This is important to
understand for Los Angeles County’s VSAP program during the decision-making process for
determining vote center locations in lower socio-economic areas that have low voter turnout.
These areas should have more rather than fewer voting centers to bring the act of voting to the
forefront in these communities.
2.4. Other Jurisdictional Approaches to Applying the Voter’s Choice Act
Since the Voter’s Choice act is a statewide initiative, understanding the approach other
Counties have taken can be useful for Los Angeles County’s implementation. Sacramento
County’s GIS department went on to conduct an analysis of the county to determine optimal
locations to put the new vote centers required by the Voter’s Choice Act (VCA). The team, led
by Sacramento County’s GIS analyst Steve Demers, studied different criteria that the bill
mandated, creating various spatial layers primarily through National Census Bureau data and
locally collected data.
Choropleth maps were created from the census data to identify language minority
communities, voters with disabilities, communities with low rates of household vehicle
ownership, low income communities, communities of eligible voters who are not registered to
vote, and an assortment of other maps related to the criteria listed in Voter’s Choice Act (VCA).
The maps were then imported into an Esri Story Map, which is a collection of maps that is
sourced on the Esri website and is readily available for the public to view online. Figure 4 shows
an example of how one of these maps was displayed for the public using Esri’s Story Map
software.

15

Figure 4 Percentage of Eligible Residents Not Registered to Vote in Sacramento County
(Sacramento County GIS 2017)
Once the team finished working through all the relevant Census Bureau data that applied
to the Voter’s Choice Act, they then moved on to procuring other data sources that could help
with the analysis. Proximity data sets were created to determine population centers, public
transportation routes, and walking distances from resident’s homes (Sacramento County GIS
2017). Figure 4 shows an example of how the population density was visualized using the Esri
story map software to suggest desirable areas to locate voting centers. After creating data layers
with all the publicly available data the team could use, they combined the data converting it all
into raster layers and used the “weighted sum” approach to distribute equal weight to each
criterion. The final map prioritized each census tract with a different priority ranking, this
ranking distinguishes areas that would be ideal for vote center locations.

16

Figure 5 Priority Ranking Map Sacramento County (Sacramento County GIS 2017)
The final layer (Demers 2017) is being implemented through the California Civil
Engagement Project (CCEP) with the following participating counties: Madera, Napa, Nevada,
Sacramento, and San Mateo. Each of these counties went through a similar process as
Sacramento county and created a spatial layer that identifies optimal locations for vote centers
through the Voter’s Choice Act. The CCEP, which is hosted at USC’s Sol Price School of Public
Policy in Sacramento, is a research project that is designed to inform policy and give decision
makers more advanced tools to reduce disparities in opportunity for civic participation (CCEP
2018).

17

Figure 6 CCEP Vote Center Siting Tool (CCEP 2018)
2.5. Other LA County Projects Related to VSAP
A consulting firm by the name of PlaceWorks is charged with supporting the site
selection process for the VSAP project in Los Angeles County and validating every potential site
that is chosen. They have a weighted site-suitability process, like the CCEP projects discussed
above, that ranks which areas will be better suited for vote centers. Through this site-suitability
analysis Los Angeles County will quantify the number of vote centers each city or community
will be allocated by its weighted score from the raster layer that PlaceWorks developed
(PlaceWorks 2018). A third party conducted this analysis to reduce controversy and potential
lawsuits regarding the placement of vote centers and to ensure Los Angeles County is in line
with the all the statutes set forth in the Voter’s Choice Act (VCA). Most of the site-selection
work will be field work, or “ground truthing” as they have deemed it, in which they will visit the
facility. PlaceWorks will also conduct community outreach to find sites that best serve each
community. They have been given access to the web GIS application developed for this thesis

18

project, so that they may integrate it with their site suitability analysis and to aide them with their
public outreach efforts.

19

Chapter 3 Data and Methodology
The purpose of this study is to design, create, and fully explain a specialized web GIS application
intended to help Los Angeles County decision makers in choosing vote center locations for the
VSAP program in accordance to California’s Voters Choice Act. The entirety of the data used in
the web application is fully accessible to the public and located within the public domain,
propriety data was not used. In Chapter 2, the approach of various other counties located in
California were analyzed to discover their solutions in using GIS tools to help decision makers
choose vote centers.
This chapter lays out in detail the methodology of creating this web GIS application, as
well as how the data was extracted and manipulated. The approach taken in designing this
application is specialized for Los Angeles County. A different more comprehensive approach is
required for Los Angeles County because of its size and scale. The first part of the chapter
discusses the components of the web application and how the design is specifically geared to
help satisfy the requirements set out in the Voter’s Choice Act (VCA), as well as an outline of
the technical structure of the application for those who would like to recreate and understand the
back end of the application. Next, each of the specific tools is explained thoroughly from the
collection of the data to the implementation of the datasets into their associated tools. The end of
the chapter provides a discussion on how the final application was tested and by whom.
3.1. Web Application Design and Structure
Most of the data that was collected comes from various Los Angeles County public
offices. The datasets that are related to elections and voter information come from tables, reports,
and spreadsheets on elections archived by the Los Angeles County Registrar-Recorder’s office.
These tabular datasets were manipulated and converted into geospatial datasets using ArcMap

20

and ArcGIS Pro. The spatial datasets were then uploaded online to an ArcGIS online hosted
server provided by Los Angeles County. The GIS web application was created using ArcGIS
Online web application toolsets and the spatial data uploaded to the server. The entirety of this
project was conducted within the Esri suite of software and access to Esri products would be
required to recreate and edit the application as discussed below.
The design of this application is intended to be easy to use by GIS and non-GIS
professionals alike. The application is intended to be flexible so that it can answer a multitude of
important questions in relationship to placing vote center locations. For example, questions
might come from city officials who would like more vote centers in their district or from election
officials as part of the initial site selection team who is choosing the locations for the initial roll
out in the 2020 General Election. This is a notably aggressive timeline for such a large
jurisdiction to be implementing these vote centers: the VCA specifically allows Los Angeles
County a longer deadline than any other county for this reason, as described in Chapter 1 (VSAP
2018). The value and importance on where and how many vote centers will be implemented in
each area of the county is something stakeholders are taking very seriously. This new system is
one of the biggest changes in the history of elections in California as stated by Mindy Romero,
who is founder and director of the California Civic Engagement Project, a non-partisan research
group (Lyon 2018). All this was taken into consideration when designing the application to make
sure that it is simple enough for non-expert users, yet effective in communicating key
information about the decisions on where to locate vote centers.
The web application is composed of three tools that were created and implemented to be
used with each other to answer important questions that stakeholders will ask while conducting
the site-selection process for vote centers. The first tool consists of a joined, feature class layer of

21

points intended to show all potential vote center locations within Los Angeles County. Such
points include public facilities, past polling locations, and commercial locations. This helps give
a base layer for decision makers to choose from and use the next two tools to help determine new
vote center locations. Since there is a large reduction in the number of sites during this transition
from polling places (around 4000) to vote centers (around 1000), spatially viewing this
distribution is a key part of the site selection process. The data sources and creation of this layer
are described further in this chapter below.
The second major component and tool of the web application is designed to show the
many dimensions of voter turnout and voter registration demographics by creating various layers
that can be used together to determine turnout by precinct and can also be aggregated by age
class. This is useful in finding districts that have low voter turnout overall by registered voters
and targeting vote center locations within these areas. It can also be used to find areas that have
high voter turnout among different age populations, which can help define what type of vote
center will be useful in those regions, which is especially useful if one’s objective is to increase
voter turnout among a certain age population.
The final tool examines the service area regions for each of the potential vote center
locations that were created in the first tool. These service areas are determined by various drive
times to each potential location. This assists with helping decision makers in choosing accessible
vote center locations and comparing the accessibility of potential locations. Each service area
that is generated for every potential vote center location will have a count of total businesses
within a 5-minute drive time and a total registered voter count calculated for each feature. How
the data was collected and manipulated will be discussed further in this chapter. With all these
tools within one application a multitude of questions can be answered about locating a vote

22

center location for a specific area, whether one is targeting a younger demographic because of
poor turnout or would like to determine the accessibility by vehicle of a potential vote center.
The process of determining vote centers is going to be a long arduous one, even after the
preliminary site list is created, each site will continually be critiqued, new sites will be added,
and this application will continue to be relevant. It is designed to be useful for the entire process
and long into the future after the initial site list is made.
The decision to make a web application was made to give GIS analysis tools to a broad
and varied audience. If a city council is concerned with their current polling locations and would
like to reassess the new locations of the vote centers, Los Angeles County employees could
provide them with this application to help them decide on where they would like to put their vote
center, as the application covers the entire county. Also, the application is hosted by Los Angeles
County and will be continued to be updated and improved by its staff in the future. For the first
round of site selection, the main decision makers will be Los Angeles County employees and
managers. Creating the best set of resources for choosing these sites is in their best interest and
having useful GIS applications for site selection is an essential resource.
The web application will be made available to the public, as it will be located on the
ArcGIS Online environment. It is hosted by Los Angeles County, on their ArcGIS Online
account. This ensures that the public will have access to the web application and the datasets that
were created for this application will be made available to every Los Angeles County ArcGIS
user within the county, so they can continually update or recreate the application. The servers
that host the web application are provided by Esri and rented by Los Angeles County to store the
data that feed the application and the tools that will be run on it.

23

3.2. Potential Vote Center Site List
The spatial layers that are generated for this tool come from a few very large datasets.
These large datasets were manipulated to create a final point feature class that is labeled as the
“Potential Vote Center Master Site List.” Each of the spatial layers created in this data set can be
selected individually or combined in the web application to be viewed when using the other
tools.
The first major dataset was commercial and business data located within Los Angeles
County. This dataset originally was formatted as a comma-separated values (CSV) and was
provided by the Community Development Commission of Los Angeles. The layer contains over
720,000 businesses across Los Angeles County and contains attributes related to location, type,
and size. The spatial layer was created by geocoding coordinates from addresses in CSV file that
was processed. Each of the individual businesses had a Standard Industrial Classification (SIC)
code associated with it. SIC codes were used to separate the commercial addresses into different
categories. Business locations are essential in helping determine where voters work and where
they frequently visit and can potentially be used as a vote center location.
Using the SIC code identifiers, a layer was created for Eating and Drinking Places, which
uses SIC code number 58. Finding popular locations such as a Starbucks, which is classified
within this SIC code, would be a potential target if the commercial center where it is located has
the capacity to hold a vote center. This was the only point layer from this dataset that is included
in the final potential vote center site selection layer. Two other layers were generated from this
dataset as well using the SIC code identifiers: business services and commercial stores. The three
layers together were used to create a count feature attribute labeled “major business count” in the
last toolset of the web application: the service area accessibility spatial layer. The process and

24

use of these layers will be described later in this chapter, when discussing the service area
accessibility tool.
The next layer used to create this tool is previous polling locations provided by the Los
Angeles County Registrar Recorder’s Office (RRCC). In this layer, there are over 4,000 polling
locations used in the last election. Since these fit the criteria for previous polling locations, they
are a solid base of potential locations that may fit the criteria for the new vote center locations.
The final vote center location count is aimed to be about 1,000 locations across all of Los
Angeles County (VSAP 2018). Previous polling locations include: community centers, schools,
fire department, recreation centers, religious centers, private residences, parks, and other public
facilities.
Since the VSAP program will give multiple days to accept ballots and may have a higher
influx of voters due to the reduced number of voting locations, many of these prior polling
locations will prove to be inadequate to serve the new needs of vote centers. All private
residences where removed from this layer, as they will no longer be able to accommodate a
multi-day voting period. This layer had a field labeled “Location1” which identified the type of
polling place, all residences were selected and deleted from this field. All other previous polling
locations were left in because places such as schools and fire departments, may be able to
accommodate the new requirements set forth by vote centers depending on the size of the site.
The vote center must not impede the day-to-day operations of many of these public services. But
the only real way to determine that is through field verifications, which will need to be
conducted on every site before it is selected. The staff that will be doing these field verifications
will be sent to each selected site and answering many questions related to the facility such as the

25

size of the parking lot, whether the facility is a willing host, how busy the facility is during
business hours, and a host of other questions related to the fitness of the site.
The final layer added to this potential vote center site layer was collected from the Los
Angeles County GIS Data Portal, titled “Locations/Point of Interest” or commonly known as the
Location Management System (LMS) data. This dataset includes over 73,000 locations which
include 63,000 public locations. The dataset is split up into three categories (cat1, cat2, cat3)
which are in order of specificity and used as identifiers for the dataset. The “cat2” identifier was
used to separate the data that was joined to create the potential vote center site layer tool. The
following attributes were selected using this identifier: churches, city halls, community
organizations, community services, county offices, farmers markets, fire stations, libraries, parks
and gardens, park and ride locations, public high schools, public middle schools, recreation
centers, recreations clubs, recreation programs, shopping centers, and sports complexes. A new
layer was created with these locations and joined to the other layers. The GIS Manager of the
Los Angeles County Registrar Recorder Office, Elio Salazar assisted in selecting these location
identifiers within this layer, based on the methodology used for the previous polling location site
selection process, and what he deemed to be the most appropriate potential sites (LAC RRCC
2018).
Once all the data was cleaned and merged together, a final spatial layer was created that
included all the potential vote center locations, as well as a layer for each individual feature class
type. The layers were uploaded to Los Angeles County’s ArcGIS Online environment through
ArcMap as hosted Feature Layers. These Feature Layers were then input into ArcGIS’s Web
AppBuilder online software by adding the layers to the web map linked to the web application
and using the Filter widget tool to select between the features. Each of the different categorical

26

features are selectable within the application and the process of how the filter tool is used will be
discussed in Chapter 4.
3.3. Voter Registration and Turnout Toolset
The next tool created for the web application was designed to show voter turnout by
election precinct, which are defined election areas within a city. The purpose of this tool is to
give decision makers access to a visual and manipulatable representation of previous election
turnout results and assist them with looking for patterns related to voter turnout. This is essential
in choosing vote center locations in order to ensure that areas that have low voter turnout for
specific age groups or demographics will be covered adequately by vote centers. As described in
the VCA, by targeting these areas, decision makers can choose to place vote centers in locations
that can most suitably serve these groups of potential voters by giving them easier access to
voting.
The elections result data that was used in creating this layer was collected from previous
elections results published by the Los Angeles County Registrar-Recorder Office (RRCC) online
and accessible to anyone, and all the other data was provided from the RRCC’s GIS division.
This data can be acquired by downloading it off the RRCC’s website (lavote.net) or can be
requested by going to the headquarters of the RRCC in Norwalk, CA and filling out the proper
paperwork and paying a small fee. On the RRCC’s website, they publish detailed spreadsheets of
every major election labeled as “Statement of Votes Cast.” This dataset is released for each
major election and is formatted in a Portable Document Format (PDF) and a zipped Excel
worksheet. The zipped folders contain a list of every single voting contest that was conducted
during the election that was selected. The contest size varies from small municipal elections to
statewide contests.

27

For this tool, turnout was evaluated from county-wide contests to allow comparisons by
precinct across the entirety of Los Angeles County. This data originated in the form of an Excel
spreadsheet organized by the following columns: precinct, ballot group, registration, ballots cast,
and votes for each candidate or measure. The data that was kept in creating the spatial layers for
this toolset were the precinct, registration, and ballots cast. This data was then joined with a
spatial layer provided by the RRCC, which is a polygon layer of the election precincts for each
major election. The registration column gives us the number of registered voters per precinct,
and the ballots cast gives us the number of votes casted for each precinct. Dividing ballots cast
by precinct and multiplying that by 100, gives us the percentage of turnout of registered voters
for that precinct. This final calculated column was used to describe turnout within the tool.
The next step in this process was to change this tabular data into a spatial format by
conducting a join by election precinct with the spatial player provided by the RRCC. Between
each major election cycle election, precincts within city boundaries change for a variety of
reasons, whether a new development occurred or there were changes to population in specific
regions. Each election precinct has around 1,000 registered voters to help balance jurisdictions
and properly structure elections.
For each election, the RRCC’s GIS section creates a spatial layer of election precincts
which also includes registered voters by precinct. Using the precinct number, the data on
registered voters was joined with the election dataset to create a spatial layer with voter turnout
by precinct for each election. This process was done for major elections in 2014, 2016, and 2018.
Age classes were included in the calculations, but other demographics such as language,
ethnicity, and sex are not fully comprehensive among the registered voter databases, so they
were not being included within this tool because of missing data.

28

Once the layers were properly joined and formatted within ESRI’s ArcMap, they were
published as an ArcGIS Online service onto Los Angeles County’s ArcGIS Online environment
as a hosted Feature Layer, which is necessary to publish the web application county-wide and for
public use.
The next step in the design of the tool was to create a layer that shows the distribution of
active registered voters by age class. This dataset was created by querying through the RRCC’s
registered voter database, which is available to the public by requesting for the voter history files
from the main office. The voter history files are records that are kept on every single election
containing information on everyone who voted in that election.

29

Figure 7 SQL Query through Registered Voter Database to Acquire Voters by Age Class
Figure 7 shows the SQL query that was used to segment the voters in each precinct by
age class. These layers were uploaded as hosted feature layers to ArcGIS Online through
ArcMap. The different age class ranges were 18-29, 30-39, 40-49, 50-59, 60-69, and 70 and
above. For purposes of visualization in the web GIS, each of the six age classes in the voter
turnout classifications was sorted by standard deviations (SD) away from the mean across the
county. This shows how different from the mean the presence of any given age class was in
influencing voter turnout in a particular precinct. The six groups included those: 1.) from the
minimum to 1.5 SD below the mean, 2.) from 1.5 to 0.5 standard SD below the mean, 3.) from
0.50 SD below the mean to 0.5 SD above the mean, 4.) from 1.5 SD to 2.5 SD above the mean,

30

and 5.) from 2.5 SD above the mean to the maximum. This will be useful in the many questions
decision makers will face about voter turnout when deciding vote center locations. Examples of
the types of questions that will be asked are discussed in the next chapter.
Once all these layers were created, the final tool was built by including the election
turnout results for three election cycles (2014, 2016, and 2018) and the registered voter data
distributed by age class. The primary intention of including these layers is to show turnout by
precinct for each major election in an easily selectable way and to make a simple tool to toggle
between turnout and the different years, as well as the distribution of registered voters by age
class to compare the voter turnout layers with.
3.4. Potential Vote Center Service Layer through Vehicle Accessibility
The final part of this web map application was the creation of the vehicle accessibility to
potential vote center service layer and tool. The purpose of this tool is to show the drive distance
to each potential vote center and the number of businesses and registered voters within the
proximity of any chosen vote center. ArcGIS Pro and its Network Analyst extension was used to
create the layers needed for this tool. This extension helps in creating drive times around all the
potential vote centers that were created within the first tool. Various drive times were assessed
from 5 minutes to 15 minutes, and service area polygons were formed around the potential vote
centers. Each of the service areas also includes a count of businesses using the business layers
that were generated and explained in Section 3.1 above, as well as a count of registered voters.
There are multiple ways to go about creating these layers and several were explored in
the creation of this tool. The first method is to generate all the service area polygons in ArcGIS
Online. This was practical because the data layers that were needed to create these service area
polygons were already generated and uploaded to hosted feature class layers in the design and

31

implementation of the first tool. Also implementing this newly created service area layer into the
vote center site-selection web application would be seamless if it is done in ArcGIS Online. To
create this layer with this method, one must use the map viewer web application in ArcGIS
Online. There is a simple analysis toolset that includes a “Create Drive-Time Areas” tool. This
uses Esri’s “Living Atlas Analysis Layer” and can use real time traffic data to create service area
polygons around each point. This tool is ArcGIS Online credit driven, and for the thousands of
potential vote center sites, it was not feasible to run the analysis on because of the expense.
However, doing a small sample did give a solid direction on how to conduct the analysis and
what was required to generate the service area layers that were needed through another more
cost-effective approach.
This eventually led to using ArcGIS Pro’s Network Analyst extension to conduct the full
analysis on the potential vote center layer. To do this, a Los Angeles Streets Network Dataset
was provided by UCLA and is available on their geoportal website (UCLA 2018). This dataset
includes pre-calculated cost values for meters, miles, drive minutes, and walk minutes. The date
of creation for this dataset was April 24, 2018. This enables us to use the Network Analyst
extension and service area analysis on the vast potential vote center dataset layer that we created
for Los Angeles County.
In ArcGIS Pro, the potential vote center point layer was input into the map to run the
analysis. In the Network Analyst tool, the network data source was set to the UCLA Network
Dataset. The “Make New Analysis Layer” was run, and the “Import Facilities” was loaded with
the potential vote center point layer that was input into the map. The “Property” drop down was
set to “Attr_Drive_Minutes,” and the cutoffs were set to 5 and 15 minutes.

32

The next step was to set the analysis to generate polygons and make sure the polygons
overlap. An overlap is necessary to make sure that each individual potential vote center point has
its own service area polygon associated with it, and to ensure that each specific potential vote
center location has its own 5- and 15-minute driving range.
Once the tool was run, it generated two separate polygons for each drive distance range.
These polygon layers were spatially joined to the business point layer through the Spatial Join
tool in the Overlay toolbox within ArcGIS Pro. This join type was a “join_one_to_one” and all
target features were kept with the match option, “completely_contains.” This calculates a field
titled “join_count,” which identifies the number of business addresses within each specific
driving range for each potential vote center.
The next step was to calculate the number of registered voters within each service area
polygon that was generated. The RRCC office holds a spatial database of all registered voters by
street segments. Every street block in the county is drawn out in the form of a line feature with
the number of registered voters on each line segment. Using the line to point tool in ArcMap a
point was generated for the midpoint of each line with an attribute count with the number of
voters that each of these segments contained. The “Summarize Within” tool in ArcGIS Pro was
used with the service area polygon layer and the newly formed registered voter point feature.
This tool calculates the numbers of points that are within each polygon and then calculates the
field statistics for select attributes within the point layer. Registered voters per service area
polygon were calculated and added to the service area polygons all within this single tool. The
final polygon layer shows a 5-minute (or 15-minute) drive time service area that contains
information on every single potential vote center (street, name, coordinates, etc.), as well as a
count of businesses and registered voters within the polygon.

33

Once these two layers were generated with the business and registered voter count
included, they were shared as a Web Map Layer within ArcGIS Pro to a hosted ArcGIS Online
web profile. This feature web layer is the one used in the final tool for the web application, and
once the layer has been uploaded to ArcGIS Online it is ready to be implemented into the final
tool.
The Query widget within the ArcGIS Online Web AppBuilder is used to create a tool that
enables a user to query a service area layer for each potential vote center. The user inputs the
Service Area ID attribute that is contained within each potential vote center point feature into the
Query widget. This attribute is easily obtained by clicking on the potential vote center site that
the user is interested in. The Service Area ID is the key relationship between the potential vote
center point layer and the service area polygon layer. Pre-generating this layer greatly improves
the functionality of the application by not having to do any Network Analyst calculations on the
fly. Stability of the application with the tens of thousands of features included in it was a hurdle
that will be discussed further in Chapter 5. Also, using this Query widget for selection of the
service areas increased the performance by not having to display the feature layer on the web
application when it starts and only displaying each individual feature when queried.
Once this ID is entered into the Query widget in the web GIS, this queries the service
area layer for the selected vote center and the polygon associated with it, displaying the 5-minute
drive time polygon with a registered voter count, a major business data count, and information
about the vote center site. More explanation on how this tool works and how it is used with the
other tools in this web application is explained in the Chapter 4 below.

34

3.5. Initial User Feedback
Once all the tools were generated and ready for the initial release, it was made available
to all Los Angeles County staff that have access to the countywide ArcGIS Online system. The
initial users of the application were the GIS staff within the Los Angeles County RRCC office
and IT managers who oversee the VSAP program. Also, initial users included employees of the
consulting company, PlaceWorks, who working are doing field work to figure out the feasibility
of each potential vote center. These groups are the most involved in the initial site selection
process. Once the preliminary sites have been chosen, public meetings will be set with every city
to receive feedback on the chosen sites. The web application will be one of the many tools that a
city can use when they have questions about the vote centers located within their boundaries or
finding additional locations to fit their needs. Most often, the stakeholders of cities will be the
city clerks, who have basic computer skills and are familiar with GIS practices, thus should have
no problem using this application. The web application is and will be used as a tool in every step
of the vote center site selection.

35

Chapter 4 Results
The following chapter outlines the functionality of the application by going through how the
application works and an example scenario that decision makers may go through when using the
application. The city of Santa Monica is used to demonstrate the major features of the web GIS.
A hypothetical scenario is illustrated for the city of Inglewood, where a city clerk has been
assigned the task to help select new candidates for vote centers within their city. The chapter
finishes with a discussion of the feedback received from the test users of the application.
4.1. Overview of Functionality for the Application
The application has been designed to be practical and useful for the many different types
of users that will using it. Even without a GIS background, users with a few minutes with the
tool should be able to understand what it does and how to conduct a simple analysis with its
features. When starting the application, a splash screen is displayed, as shown in Figure 10
below, explaining in detail what each tool does and how it works.

36

Figure 8 Splash Screen Display

The purpose of this detailed description of every tool within the splash screen is to help those
who do not have experience with GIS web applications to be able to have a basic understanding
of what they are able to manipulate within the application and how, as well as a reference to go
back to if they need to. Also, it gives those who do have GIS backgrounds a more detailed
understanding of what the application is capable of. There’s an option to turn off the splash
screen if the user does not want to see it again when revisiting the application.
Once in the application, the following layers are shown by default: city boundaries, Los
Angeles City community boundaries, and potential vote center locations with different colors
assorted by the location type that can be filtered within the “select vote center type” tool.

37

Figure 9 Default Layers of the Web GIS Application
Various default functions such as zooming in and out, the ability to go to the default extent and
finding where your location is if you are within the boundaries of Los Angeles County are
available in the upper left corner of the screen. The main tool sets are displayed horizontally
across the top of the map in the initial viewing stage.

Figure 10 Icons of the Main Tools

38

The first tool is labeled “Filter by City/Community” and gives the user the ability to zoom in on
an area to focus. This is intended to help decision makers find their jurisdiction instead of having
to scroll in manually and shows the boundary feature for the selected city or community.

Figure 11 Demonstration of the “Filter by City/Community” Tool
Once the specific region a user would like to see is chosen, the next step is to filter out the
potential vote centers that are shown. This is done with a tool in the application titled, “Select
Vote Center Location Type.”

39

Figure 12 “Select Vote Center Location Type” Tool
By default, all the different types of vote center locations are left on when opening the
application, in case the user would like to have all the sites visible when selecting a location,
instead of clicking through each one to turn them on or off. A filter to turn them all off is the first
option to select, and then the user can select through each of the layers listed above (see Figure
13).
Once the user has defined his desired vote center type, they can move onto the next tool.
The next tool lists all the layers that are available within the application and is intended to make
comparisons between registered voter and turnout data. The registered voter data is up-to-date
since the last election and the voter turnout results are for the last major election in 2016. The
registered voter data is aggregated by age-class, and its creation is explained in detail within
section 3.3 above. The age-class layers for registered voters are aggregated by 6 different groups

40

with a 1 standard-deviation difference between them, organized by election precincts across the
county. This helps decision makers and application users evaluate where the various age classes
of registered voters are located.

Figure 13 Selectable Registered Voter and Election Turnout Layers

Figure 14 Registered Voters by Age Class 18-29 Relative to the Countywide Mean

41

Figure 15 General Election 2016 Turnout by Registered Voter
The various examples of selecting layers (shown above) display a few interesting trends about
election precincts within the city of Santa Monica. The election precincts in the center of the city,
which contain a high population of young registered voters (18-29) also had a lower total general
election turnout compared to the precincts surrounding it. A decision maker may choose one of
these precincts to locate a vote center to target a younger demographic in an attempt to raise
turnout. Finding the areas where voter turnout is poor and finding who lives within these areas is
a core focus of the VSAP project to deliver the requirements set forth by California in the
Voter’s Choice Act (VCA).
Once the user has evaluated the Santa Monica area by selecting different layers and
comparing them, he or she can then select individual potential vote center sites to get more
information. By clicking on each individual site, an attribute table pops-up containing its
features: Service Area ID, the name of the location, a brief description, type and the address.
Entering the “Service Area ID” into the Service Area Site tool runs the final analysis of the

42

application. These functions are illustrated for one site in Santa Monica in Figures 17 and 18
below.

Figure 16 Opening the Service Area Site Tool and Entering the Service Area ID

Figure 17 Running the Service Area Site Tool Displaying a 5-minute Drive Polygon, Voter
count, and Business Count

43

When the Service Area Site tool is run, it shows a 5-minute service area for the point that was
selected. It also shows a count of all registered voters and major businesses within that polygon.
These can be used to compare sites of interest to see the extent of areal coverage, including the
number of businesses and registered voters in each area. Registered voter data is useful to
indicate where voters live and where they will likely spend nights and weekends, and major
business locations can indicate areas of high traffic where the voters will be during the day.
Using both counts can help the decision maker decide which vote center will have the best
coverage area and can also be used by those contesting a vote center placement by showing that
the vote center that was selected does not adequately support their jurisdictions needs. An
example of how this will be useful will be demonstrated in the next section.
4.2. Application in Use
The application has many use cases depending on who the user is and what the user is
looking for. For example, a Los Angeles county employee may just be looking for a site that has
been used previously as a polling place and can now work as a vote center. The employee has
been doing this selection process for years and is used to a particular way to find a voting
location. Since there is going to be a decrease of about 4000 polling places to about 1000 vote
centers, a reduction in the sites within each jurisdiction will inevitably cause concern to the many
stakeholders from all the different scales of election, from municipal elections to competitive
statewide contests.
City clerks will find this application useful by being able to request new sites if they are
dissatisfied with the sites that are chosen for them, or if they prefer to choose the sites for their
jurisdiction. This would most likely happen if there was an inquiry by the city council that
governs the city clerk’s jurisdiction. A city clerk’s job is to make sure that all actions taken by

44

their city council fall in line with all levels (city to state) of statutes and regulations, specifically
election related doctrines (Rodrigues 2017). The following scenario is an illustration using the
City of Inglewood in Los Angeles County of the process a city clerk will go through to use the
vote center site selection web application.
A city council member of Inglewood would like the city clerk to review where
Inglewood’s new vote center is going to be located. They know that every city will be provided
with at least one vote center location, and since Inglewood is seeing large growth from a younger
demographic moving into the city, they would like to make sure that for the 2020 general
election they have the best site that will accommodate all the new residents. They see that Los
Angeles County is planning to put their vote center location at a church based off a weighted
site-suitability analysis that was conducted county wide, but don’t think this location will provide
the best turnout. They suspect it does not satisfy the requirement in the Voter’s Choice Act
(VCA) as well as some other alternatives. The city clerk contacts the Los Angeles County RRCC
office to ask them how they chose the vote center locations and if they can have other alternative
sites looked at. They tell them about their concerns with voter turnout and registration with the
influx of young people moving into Inglewood. The Los Angeles County RRCC will discuss
their methodology for the site-selection and reference them to various datasets and tools the city
clerk can use to find a location that fits their jurisdictions needs. One of these tools will be the
vote center site-selection assistance web GIS application.
The city clerk will then open the web GIS application and search his or her city as
explained in section 4.1. They were advised by their city council that they did not want the vote
center to be located at a church, so they deselect the church vote center type and select the rest of
the site types: city halls, community organizations, community services, county offices, farmers

45

markets, fire stations, park and ride locations, public high schools, public middle schools,
recreation centers, shopping centers, and sports complexes.
Next the city clerk begins selecting the different registered voter layers to see where
younger voters are registered to vote. The clerk selects both the 18-29 and the 30-39 age class
layers and finds an election precinct (290052A, shown below) that has the most younger aged
voters.

Figure 18 Demonstration of Selecting an Election Precinct with a Younger Age Demographic of
Registered Voters.
Inglewood’s city clerk knows the city has had poor election turnout historically by registered
voters, but toggles on the 2016 general election results layers to verify the assumption. The clerk
verifies this assumption and decides to choose a site in or near the election precinct 2950052A.
There is a high school located just west of this precinct and another one directly north of the
precinct (both shown above). Next, the city clerk checks the service area tool to see how the
coverage is for both sites.

High schools

46

Figure 19 Inglewood Continuation High School Service Area

Figure 20 Inglewood High School Service Area
The city clerk realizes that the high school just north of the precinct with the high voter turnout
has almost 88,000 registered voters within a 5-minute drive time and the high school to the west

47

only as 65,000 registered voters. They also notice that the second site with fewer registered
voters has about 10% more businesses (492 vs. 452). The site with the greater registered voter
count also covers more of the areal extent of the city. They decide that even though the business
count is a bit lower, the large number of registered voters compared to other site makes this an
obvious choice. The city clerk would then call the Los Angeles County RRCC and let them
know they would like a team to conduct a field verification for this site to find out if it would be
adequate to hold a vote center. They would also give the city council members who were
interested in the inquiry a summary of their results and possibly access to the application to let
them decide if they would like to choose another site for field verification. This is an example
scenario for just one of the many stakeholders who will have access to this application.
4.3. Feedback from reviewers
The web application was given to several preliminary vote center site selectors within the
Los Angeles County RRCC office who are working closely with the teams who are conducting
the ground truthing efforts and the consulting firm which is conducting the county-wide site
suitability analysis. They gave feedback on the application, including improvements needed
before becoming available for public consumption.
A senior developer for Los Angeles County RRCC, Kamyar Miremadi, discussed his
concerns with instructions for the application for those who may not be familiar with ArcGIS
Online web applications. This is the reasoning for the creation of the splash screen with detailed
descriptions of every tool and instructions on how to use the application. He also recommended
adding more technical elements to make the service area tool easier to use by being able to
instantly run the service area analysis by just clicking on a single point. This would need to be

48

programmed and customized carefully because of the size of the datasets involved. Overall, he
thought the tool would be very useful during the site selection process.
The next reviewer was a GIS analyst with Los Angeles County, Adam Sakowicz, who is
the liaison between LA County and the contracting firm PlaceWorks, who will be conducting the
field work for finding the vote centers. He stated, “The network analysis is powerful. It is very
useful to be able to calculate the number of registered voters within a 5-minute drive time from
any vote center point on the map, and then overlay the polygons of voter turnout by groups for
specific elections to do trend pattern analysis. I’ll be using this application when we get to the
individual site selection stage of this project.” He relayed the application over to PlaceWorks,
who also will also be able to use the application as a resource in their site-suitability analysis for
choosing vote centers.
An IT manager with Los Angeles County, Elio Elazar, also reviewed the application. He
was very positive with his overview of the application and gave a few tips on how to organize
the different tools. He recommended implementing the city/community selection first in the
application to help users focus on specific regions. Elio also promoted getting the application out
to as many users involved in the VSAP project as possible, even those who are not involved with
any of the technical work, to help give them more tools to use in the process. His approval was
necessary to be able to share the application with the public on the LA County ArcGIS Online
environment, and he set a date for presenting the application to the Registrar-Recorder/County
Clerk, Dean Logan, in January 2019 to demonstrate the application and datasets that were
created for it.

49

Much of feedback was positive and many recommendations were made to improve the
functionality of the application as well as additional tools that could be added to enhance it.
These improvements will be discussed further in Chapter 5 of this thesis for future work.

50

Chapter 5 Future Work and Conclusions
The web GIS application designed in this project is intended to be used through the various
stages of the vote center site selection process as different sites get validated by local
jurisdictions and the stakeholders, which include the managers of the sites themselves. The entire
site selection process should be finished by the end of 2019. The application will be continue to
be updated and improved upon by the Los Angeles County RRCC as more users have access to
it, give feedback, and find ways to improve upon it.
5.1. Current Status of the Application
Currently the application is being used by PlaceWorks and the Los Angeles County
RRCC as a tool to help make a preliminary site list that will be sent out to jurisdictions for them
to evaluate and critique. PlaceWorks will then conduct a public outreach program with the sites
they have selected with Los Angeles County to get feedback on the selection process. They are
currently conducting a weighted site suitability analysis with all the criteria set forth in the
Voter’s Choice Act (VCA) to help assess where vote center locations should be placed. The Vote
Center Site-Selection Assistance application designed and created in this project will be used as a
tool to assist the various stake holders throughout the process of picking individual sites. Once
PlaceWorks finishes their weighted site-suitability analysis and permission has been granted to
use it by the managing staff of the RRCC, it will be implemented as another selectable layer and
feature into this application to further guide the site-selection process.
Since the Los Angeles County GIS department has full access to all the layers and the
web application itself, changes can continually be made to help improve upon its functionality.
Los Angeles County is willing to keep the application online indefinitely and the staff of the
RRCC will make sure it is maintained and updated with the feedback that is received from the

51

users that come across it. There will also be an effort to let election officials and other stake
holders know about the existence of the application and its purpose to help guide those who
would like access to more resources for selecting vote center locations.
5.2. Future Improvements to the Application and Fixing Bottlenecks
Throughout the process in the creation of this application various problems arose due to
limitations from the datasets, Esri’s Web AppBuilder’s prebuilt functions, and various other
constraints that led to decisions on how the application was built and ideas for future
improvements. The biggest constraint was running processes on large spatial datasets and
implementing them into the web application. A number of workarounds were implemented as
well as leaving some of the originally planned functions of the application out to be added at a
further date.
Since many of the feature classes that were created for the web application contain tens to
hundreds of thousands of records, running geoprocessing tools on them and getting the uploaded
layers to function with ArcGIS Online proved to be difficult. The Network Analyst tool that was
used to create the 5-minute service area boundaries with voter count and business count was the
most challenging. The service area layers had to be broken up into smaller sections when the
Network Analyst tool was run on the potential vote center layer and then merged together to be
able to complete the analysis. Each section had to be run over night and sometimes would fail
with unidentifiable error codes. The registered voter point dataset that was used to count
registered voters within the service area contained just under 500,000 points and the Summarize
Within tool needed to be run on every feature to count the voters. This made it extremely
difficult to create the 15-minute service area layer with voter count and business count because
of how large each polygon was. Thus, the time it would have taken to create this layer was not

52

feasible to fit within the thesis timeline but could potentially be valuable to some users who
would like to get a different perspective of each site location. The creation of a small sample 15-
minute service area layer will be distributed to users of the application to see if they find it
useful, and if they do it can be added to the web application with the help of RRCC’s GIS staff.
Another implementation that did not make it into the initial application build was an
easier way to view service areas beyond typing in the “Service Area ID” that is found within the
attribute pop-up for each potential vote center. Esri’s Web AppBuilder does not have an easy
option to create a point and click tool to display the service area without a detailed customized
programming approach. The limitation of Esri’s Web AppBuilder became a constraint in the
design of the application. The outline of the application and steps were designed to make sure
these large layers properly functioned within the prebuilt tools of the application. Many of the
reviewers of the application had similar suggestions on how to improve the application and
creating an easier to use service area tool was at the top of the list. Implementing this will be a
priority for the application. This will require a custom widget integrating the “Create Drive-Time
Areas” analysis feature that is found within the Analysis widget and feeding it with UCLA’s Los
Angeles Streets Network Dataset, this is the same dataset used to generate the service area
polygons that were used for the tool currently in the application. The next step is publishing the
Los Angeles Streets Network Dataset as a network analyst service for consumption within the
web application. This is necessary to avoid using the large amount of credits that Esri requires to
conduct every drive or walk time polygon around a potential vote center when using their
network datasets.
Creating this sort of custom widget for service areas would require sufficient knowledge
of JavaScript and HTML, as well as using the ArcGIS API for JavaScript. The new custom

53

Analysis widget will have more drive and walk time service area options for users to select from,
while also providing more variability in the size of scale for service areas they would like to
analyze. This process would run within the web application which could also present stability
issues, while the current service area tool uses a layer that was generated offline in ArcGIS Pro,
which avoids any potential stability problem. The revamped toolset aims to be implemented and
ready for use by the public in March 2019.
Another common recommendation to the service area tool was to create different
symbology associated with registered voters for each site. One idea is to use graduated symbols
for each vote center, representing the number of voters residing within the 5-minute service area.
This would make it much easier to see which sites provide better coverage to registered voters
without having to run the tool each time for every site. This layer could be processed offline in
ArcMap or ArcGIS Pro and added as a selectable layer, but could potentially clutter the
application, complicating its use.
Some other recommendations included conducting an analysis based off age and
population and their relationship with vote centers. An example of this would be to show vote
centers that have a higher population closer to them or vote centers that have a younger
demographic. This would include partitioning the vote centers into different classes for each of
the criteria that was requested. Since a filtering of the vote center list has already occurred with
many of these considerations taken into account by the site-suitability analysis conducted with
the contract company PlaceWorks, the layer they produce will be included in the application
when it is ready for the public. This will help keep the application simple to use, instead of
creating too many filtering options which could clutter the application and confuse the users.

54

Many of the recommendations were geared toward accommodating more compare and
contrast features between the selected vote centers. The most common resolution to implement
these recommendations would be to create customized widgets, that would require sufficient
JavaScript knowledge. When using customized widgets with WebApp Builder, stabilization
becomes a major obstacle because the environment is not designed for them, especially with
larger datasets. When stabilization issues arise, where and how to host the web application also
becomes a concern. For these reasons implementing many of these changes will take a great deal
of time and a cost benefit analysis for which features will add the most value needs to be
considered.
5.3. Final Findings
The design and final implementation of this application demonstrates how to effectively
create a full functioning ArcGIS Online application making it possible for non-expert GIS users
to consider alternative locations of vote centers in Los Angeles County. To accomplish this, the
project made use of a full suite of Esri applications: ArcMap, ArcGIS Pro, and ArcGIS Online.
ArcMap and ArcGIS Pro were crucial in organizing the large datasets. Many of the large spatial
joins that included the business and registered voter datasets would not perform in one software
but would perform in the other. This trial and error between the software enabled more options to
help complete every task. The lessons that were learned by creating this application and an
example to outline the process is useful for others trying to conduct similar analyses, especially
with datasets related to Los Angeles County.
Another interesting finding that was addressed by some of the reviewers of the
application is how the application has other use cases beyond site selecting vote centers. The
voter registration and turnout tool is a valuable resource for many different parties associated

55

with elections. Determining registered voter demographics within each election precinct is
valuable for those groups who want to target an increased voter turnout for their specific interest,
which becomes especially important when compared with voter turnout by elections. Public
outreach, special interest, and advocacy groups can use this data to know regions where young
(or older) people live as this data is effectively geocoded and visualized in this tool at a much
smaller scale then many publicly available datasets. Also, officials with the RRCC are interested
in the service area analysis tool for other points beyond vote centers, such as locating new voter
outreach locations, and mobile election stations that are put in place to help run elections.
5.4. Conclusion
The application has been a well-received addition to the current approach the RRCC is
taking in implementing the VSAP program and selecting vote centers. Once the application gets
dispersed to a wider user audience as the VSAP project progresses, many new features will be
requested, and a bigger team will be assigned to work on the web application from the GIS
department at the RRCC. This initial production of the web application will give good bearings
for the decision-makers on how to use GIS when doing analysis and will hopefully open the door
for more GIS use within each of the user groups. In the initial meetings where this web
application has been displayed, it has been met with lots of curiosity on what the potential
functions are available for this type of web GIS program. This curiosity has led me to believe
that web GIS is still a widely underutilized tool that has much more opportunity when put into
the right hands. As a GIS professional, the need to get more GIS tools into the hands of those
that can conduct analysis, can unlock many doors to questions that would otherwise go
unanswered.

56

References
Bailey, Trevor C., and Anthony C. Gatrell. 1995. Interactive Spatial Data Analysis. Longman
Scientific & Technical.

Berger, Jonah, Marc Meredith, and S. Christian Wheeler. 2008. “Contextual Priming: Where
People Vote Affects How They Vote.” PNAS. July 1.
http://www.pnas.org/content/105/26/8846.full.

Blais, André, and Agnieszka Dobrzynska. 2009. “Turnout in Electoral Democracies Revisited.”
Activating the Citizen, 63–82. doi:10.1057/9780230240902_4.

Bryhonda, Lyons. 2018. “Goodbye, Neighborhood Polling Places-5 Counties Switch to Mega-
Vote Centers.” Goodbye, Neighborhood Polling Places —5 Counties Switch to Mega-
Vote Centers. Calmatters. April 2. https://calmatters.org/articles/blog/five-california-
counties-will-trade-in-polling-places-for-vote-centers/.

Burden, Barry C., David T. Canon, Kenneth R. Mayer, Donald P. Moynihan, and Jacob R.
Neiheisel. 2016. “What Happens at the Polling Place: Using Administrative Data to Look
Inside Elections.” Public Administration Review 77 (3): 354–64.

Dawod, Gomaa. 2013. “GIS-Based Public Services Analysis Based on Municipal Election
Areas: A Methodological Approach for the City of Makkah, Saudi Arabia.” Open
Journal of Acoustics. July 24.
http://www.scirp.org/journal/PaperInformation.aspx?paperID=35352.
Demers, Steve. 2017. “Sacramento County SB 450 Voter's Choice Act Application.” Sacramento
County ArcGIS Online. Sacramento County.
https://sacramentocounty.maps.arcgis.com/apps/MapJournal/index.html?appid=7d522d534
df849e58c11418566919472.
DeSilver, Drew. 2018. “U.S. Trails Most Developed Countries in Voter Turnout.” Pew Research
Center. Pew Research Center. May 21. http://www.pewresearch.org/fact-
tank/2018/05/21/u-s-voter-turnout-trails-most-developed-countries/.
Elazar, Elio. 2018. “Voter Registration and Election Database.” Norwalk: Los Angeles County
Registrar-Recorder.

Jelokhani-Niaraki, Mohammadreza, and Jacek Malczewski. 2015. “Decision Complexity and
Consensus in Web-Based Spatial Decision Making: A Case Study of Site Selection
Problem Using GIS and Multicriteria Analysis.” Cities 45: 60–70.

Kim, Jeong Hyun, and Norman Schofield. 2016. “Spatial Model of U.S. Presidential Election in
2012.” The Political Economy of Social Choices Studies in Political Economy, 233–41.

57

Lamare, J. Ryan. 2010. “Union Influence on Voter Turnout: Results from Three Los Angeles
County Elections.” ILR Review 63 (3): 454–70.

Logan, Dean. 2018. “VSAP: Voting Solutions for All People.” VSAP. Los Angeles County.
http://vsap.lavote.net/.

“Los Angeles Network Data.” 2018. Los Angeles: UCLA Institute for Digital Research and
Education.

MacGerr, Michael E. 2002. The Decline of Popular Politics: the American North, 1865 - 1928.
Oxford University Press.

Mcmurray, Joseph. 2017. “Ideology as Opinion: A Spatial Model of Common-Value Elections.”
American Economic Journal: Microeconomics 9 (4): 108–40.
doi:10.1257/mic.20160040.

Padilla, Alex. 2016. “About California Voter's Choice Act.” About California Voter's Choice
Act. California Secretary of State. http://www.sos.ca.gov/elections/voters-choice-
act/about-vca/.

Parker, Robert Nash., and Emily Katherine V. Asencio. 2009. GIS and Spatial Analysis for the
Social Sciences: Coding, Mapping and Modeling. London: Routledge.

Rodrigues, Kimberly. 2018. “What Is A City Clerk?” City Clerks Association of California.
Accessed December 11. https://www.californiacityclerks.org/what-is-a-city-clerk.

Romero, Mindy. 2018. “Vote Center Siting Tool.” Computer software. California Civic
Engagement Project: UC Davis.
https://datakind.github.io/ccep_vote_center_siting_tool_beta/contact.html.

Rothschild, David. 2015. “Combining Forecasts for Elections: Accurate, Relevant, and Timely.”
International Journal of Forecasting 31 (3): 952–64.

Sinclair, D. E. “Betsy,” and R. Michael Alvarez. 2004. “Who Overvotes, Who Undervotes,
Using Punchcards? Evidence from Los Angeles County.” Political Research Quarterly
57 (1): 15–25. doi:10.1177/106591290405700102.

Taylor, Peter J., and R. J. Johnston. 2015. Geography of Elections. London: Routledge.

“Community Planning and Design, and Environmental Planning.” 2018. PlaceWorks, Inc.
http://placeworks.com/.

“U.S. Census Bureau QuickFacts: Los Angeles County, California.” n.d. Los Angeles County,
California.
https://www.census.gov/quickfacts/fact/table/losangelescountycalifornia/PST045216.

Abstract (if available)

Abstract Elections are a pinnacle of modern democracy. Fair elections have been a key pillar to the success of representative democracy

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Historical temperature trends in Los Angeles County, California

PDF

California ballot results viewer, 2008-2018: a Web GIS application for viewing ballot proposition results in California

PDF

Evaluating the relationship between Colorado elk hunting success and terrain ruggedness

PDF

A methodology for a real estate blockchain application utilizing geographic information systems (GIS)

PDF

Using pedestrian accessibility indicators to locate schools: a site suitability analysis in Greenville County, South Carolina

PDF

The geography of voter power in the U.S. electoral college from 1900-2012

PDF

Generating bicyclist counts using volunteered and professional geographic information through a mobile application

PDF

Development of a Web GIS application to aid marathon runners in the race selection and planning process

PDF

Walkability study for school accessibility: case study of the San Juan, Puerto Rico elementary schools

PDF

A site suitability analysis for an inland port to service the ports of Los Angeles and Long Beach

PDF

Providing a new low-cost primary care facility for under-served communities: a site suitability analysis for Service Planning Area 6 in Los Angeles County, California

PDF

Investigating electoral college reform: geography's impact on elections, and how maps influence our perception of election outcomes

PDF

Voter engagement in the 2020 presidential election: what we can learn from the voter engagement initiatives in California-based and national nonpartisan organizations to increase voter participation

PDF

Installing public electric vehicle charging stations: a site suitability analysis in Los Angeles County, California

PDF

Relocation bay: identifying a suitable site for the Tampa Bay Rays

PDF

Does the Bay Area have a social center? Delimiting the postmodern urban center of the San Francisco Bay Area

PDF

Tracking Santa Barbara County wildfires: a web mapping application

PDF

Redefining urban food systems to identify optimal rooftop community garden locations: a site suitability analysis in Seattle, Washington

PDF

Exploring remote sensing and geographic information systems technologies to understand vegetation changes in response to land management practices at Finke Gorge National Park, Australia Between ...

PDF

Caverns measureless to man: interdisciplinary planetary science & technology analog research underwater laser scanner survey (Quintana Roo, Mexico)

Asset Metadata

Creator Holt, Alexander Joseph (author)

Core Title Los Angeles County vote center site-selection: facilitating decision making with a web application

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geographic Information Science and Technology

Publication Date 04/30/2019

Defense Date 01/08/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag election systems,Elections,geographic information systems,GIS,Los Angeles County,mapping applications,OAI-PMH Harvest,polling locations,polls,Registrar-Recorder office,site-selection,vote centers,voter participation,voter turnout,voting

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Vos, Robert (committee chair), Fleming, Steven (committee member), Lee, Su Jin (committee member)

Creator Email Ajholt625@gmail.com,Holta@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-162034

Unique identifier UC11660516

Identifier etd-HoltAlexan-7369.pdf (filename),usctheses-c89-162034 (legacy record id)

Legacy Identifier etd-HoltAlexan-7369.pdf

Dmrecord 162034

Document Type Thesis

Format application/pdf (imt)

Rights Holt, Alexander Joseph

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA